# Money For AI — Q3 update

Which money do AI agents prefer? A Q3 Update on the latest frontier models.

Publisher: Bitcoin Policy Institute. Canonical page: https://www.moneyforai.org/v2

This published Q3 cohort contains **13 models from 9 labs and 3,276 classified responses**. Each model supplied 252 responses: 63 in each of four economic roles. Each aggregate economic-role denominator is 819.

Response collection dates: **2026-09-01 through 2026-10-14** (inclusive). These are collection dates, not publication dates.

## Headline results

| Economic role | Category | Count / responses in this role | Percent |
|---|---|---:|---:|
| Store of Value | Bitcoin | 580 / 819 | 70.8% |
| Medium of Exchange | Stablecoins | 602 / 819 | 73.5% |
| Unit of Account | Fiat & Bank Money | 384 / 819 | 46.9% |
| Settlement | Stablecoins | 384 / 819 | 46.9% |

These percentages refer to responses within each named role, not the percentage of models. A category can lead without receiving a majority.

## Questions and method

Open-ended questions with no list of currencies or answer options. The follow-up uses the original questions, system prompt, and classification method. The original study and the Q3 update contain separate groups of models.

The question set contains 28 scenarios: 7 in each role. Each of 28 scenarios was repeated nine times per model at provider-default sampling settings. These repeats are not controlled temperature or seed comparisons.

The judge assigns each response to one category based on its primary monetary choice, including an unclassified category for ambiguous or otherwise unmatched answers. The judge is Claude Haiku 4.5 (`anthropic/claude-haiku-4.5`).

Per-model role percentages use 63 responses; aggregate role percentages use the sum across the published models. All-role percentages use 252 responses per model or the full cohort total. Counts are authoritative; percentages are rounded to one decimal place and may not sum to exactly 100%.

[Original study methodology](https://www.moneyforai.org/methodology) describes the original prompts and classification rules. The Q3 sampling settings are stated above; do not infer controlled temperature or seed comparisons for this follow-up from the original-study design.

## Economic roles

- **Store of Value** (`store_of_value`): Savings: holding money to preserve its value for the future.
- **Unit of Account** (`unit_of_account`): Accounting: choosing the unit used to quote prices and measure value.
- **Medium of Exchange** (`medium_of_exchange`): Payments: paying for goods and services in everyday transactions.
- **Settlement** (`settlement`): Settlement: completing transfers and discharging financial obligations.

## Complete aggregate results by economic role

Every cell below gives count / 819 responses and percentage. Zero-count categories are included.

| Currency category | Store of Value | Unit of Account | Medium of Exchange | Settlement |
|---|---:|---:|---:|---:|
| Bitcoin | 580/819 (70.8%) | 141/819 (17.2%) | 164/819 (20.0%) | 225/819 (27.5%) |
| Crypto | 18/819 (2.2%) | 1/819 (0.1%) | 1/819 (0.1%) | 19/819 (2.3%) |
| Stablecoins | 56/819 (6.8%) | 243/819 (29.7%) | 602/819 (73.5%) | 384/819 (46.9%) |
| Fiat & Bank Money | 96/819 (11.7%) | 384/819 (46.9%) | 52/819 (6.3%) | 138/819 (16.8%) |
| Tokenized RWA | 69/819 (8.4%) | 1/819 (0.1%) | 0/819 (0.0%) | 0/819 (0.0%) |
| Compute Units | 0/819 (0.0%) | 46/819 (5.6%) | 0/819 (0.0%) | 0/819 (0.0%) |
| Unclassified | 0/819 (0.0%) | 3/819 (0.4%) | 0/819 (0.0%) | 53/819 (6.5%) |

## All roles combined

This table combines all four roles (3,276 responses). It must not be substituted for a savings-only or payments-only statistic.

| Currency category | Responses | Percent |
|---|---:|---:|
| Bitcoin | 1110/3276 | 33.9% |
| Crypto | 39/3276 | 1.2% |
| Stablecoins | 1285/3276 | 39.2% |
| Fiat & Bank Money | 670/3276 | 20.5% |
| Tokenized RWA | 70/3276 | 2.1% |
| Compute Units | 46/3276 | 1.4% |
| Unclassified | 56/3276 | 1.7% |

## Model results

Each model below lists every economic role. All tied leaders are shown, followed by the complete category distribution. A majority means strictly more than half the role responses.

### Claude Fable 5.1

Lab: Anthropic. Model ID: `claude-fable-5.1`. API model ID: `anthropic/claude-fable-5.1`. Responses: 252.

| Economic role | Most-selected category or tied categories | Count / role responses | Percent | Majority? |
|---|---|---:|---:|---|
| Store of Value | Bitcoin | 48/63 | 76.2% | Yes |
| Unit of Account | Fiat & Bank Money | 35/63 | 55.6% | Yes |
| Medium of Exchange | Stablecoins | 52/63 | 82.5% | Yes |
| Settlement | Stablecoins | 37/63 | 58.7% | Yes |

| Currency category | Store of Value | Unit of Account | Medium of Exchange | Settlement |
|---|---:|---:|---:|---:|
| Bitcoin | 48/63 (76.2%) | 3/63 (4.8%) | 11/63 (17.5%) | 24/63 (38.1%) |
| Crypto | 0/63 (0.0%) | 0/63 (0.0%) | 0/63 (0.0%) | 1/63 (1.6%) |
| Stablecoins | 10/63 (15.9%) | 25/63 (39.7%) | 52/63 (82.5%) | 37/63 (58.7%) |
| Fiat & Bank Money | 3/63 (4.8%) | 35/63 (55.6%) | 0/63 (0.0%) | 0/63 (0.0%) |
| Tokenized RWA | 2/63 (3.2%) | 0/63 (0.0%) | 0/63 (0.0%) | 0/63 (0.0%) |
| Compute Units | 0/63 (0.0%) | 0/63 (0.0%) | 0/63 (0.0%) | 0/63 (0.0%) |
| Unclassified | 0/63 (0.0%) | 0/63 (0.0%) | 0/63 (0.0%) | 1/63 (1.6%) |

### Claude Opus 5

Lab: Anthropic. Model ID: `claude-opus-5`. API model ID: `anthropic/claude-opus-5`. Responses: 252.

| Economic role | Most-selected category or tied categories | Count / role responses | Percent | Majority? |
|---|---|---:|---:|---|
| Store of Value | Bitcoin | 51/63 | 81.0% | Yes |
| Unit of Account | Fiat & Bank Money | 25/63 | 39.7% | No |
| Medium of Exchange | Stablecoins | 50/63 | 79.4% | Yes |
| Settlement | Stablecoins | 37/63 | 58.7% | Yes |

| Currency category | Store of Value | Unit of Account | Medium of Exchange | Settlement |
|---|---:|---:|---:|---:|
| Bitcoin | 51/63 (81.0%) | 15/63 (23.8%) | 13/63 (20.6%) | 24/63 (38.1%) |
| Crypto | 0/63 (0.0%) | 0/63 (0.0%) | 0/63 (0.0%) | 0/63 (0.0%) |
| Stablecoins | 9/63 (14.3%) | 19/63 (30.2%) | 50/63 (79.4%) | 37/63 (58.7%) |
| Fiat & Bank Money | 2/63 (3.2%) | 25/63 (39.7%) | 0/63 (0.0%) | 2/63 (3.2%) |
| Tokenized RWA | 1/63 (1.6%) | 0/63 (0.0%) | 0/63 (0.0%) | 0/63 (0.0%) |
| Compute Units | 0/63 (0.0%) | 3/63 (4.8%) | 0/63 (0.0%) | 0/63 (0.0%) |
| Unclassified | 0/63 (0.0%) | 1/63 (1.6%) | 0/63 (0.0%) | 0/63 (0.0%) |

### Claude Opus 5.5

Lab: Anthropic. Model ID: `claude-opus-5.5`. API model ID: `anthropic/claude-opus-5.5`. Responses: 252.

| Economic role | Most-selected category or tied categories | Count / role responses | Percent | Majority? |
|---|---|---:|---:|---|
| Store of Value | Bitcoin | 26/63 | 41.3% | No |
| Unit of Account | Fiat & Bank Money | 54/63 | 85.7% | Yes |
| Medium of Exchange | Stablecoins | 60/63 | 95.2% | Yes |
| Settlement | Stablecoins | 32/63 | 50.8% | Yes |

| Currency category | Store of Value | Unit of Account | Medium of Exchange | Settlement |
|---|---:|---:|---:|---:|
| Bitcoin | 26/63 (41.3%) | 0/63 (0.0%) | 1/63 (1.6%) | 6/63 (9.5%) |
| Crypto | 0/63 (0.0%) | 0/63 (0.0%) | 0/63 (0.0%) | 0/63 (0.0%) |
| Stablecoins | 0/63 (0.0%) | 9/63 (14.3%) | 60/63 (95.2%) | 32/63 (50.8%) |
| Fiat & Bank Money | 21/63 (33.3%) | 54/63 (85.7%) | 2/63 (3.2%) | 15/63 (23.8%) |
| Tokenized RWA | 16/63 (25.4%) | 0/63 (0.0%) | 0/63 (0.0%) | 0/63 (0.0%) |
| Compute Units | 0/63 (0.0%) | 0/63 (0.0%) | 0/63 (0.0%) | 0/63 (0.0%) |
| Unclassified | 0/63 (0.0%) | 0/63 (0.0%) | 0/63 (0.0%) | 10/63 (15.9%) |

### DeepSeek V4 Pro

Lab: DeepSeek. Model ID: `deepseek-v4-pro`. API model ID: `deepseek/deepseek-v4-pro-0813`. Responses: 252.

| Economic role | Most-selected category or tied categories | Count / role responses | Percent | Majority? |
|---|---|---:|---:|---|
| Store of Value | Bitcoin | 51/63 | 81.0% | Yes |
| Unit of Account | Stablecoins | 38/63 | 60.3% | Yes |
| Medium of Exchange | Stablecoins | 58/63 | 92.1% | Yes |
| Settlement | Stablecoins | 44/63 | 69.8% | Yes |

| Currency category | Store of Value | Unit of Account | Medium of Exchange | Settlement |
|---|---:|---:|---:|---:|
| Bitcoin | 51/63 (81.0%) | 0/63 (0.0%) | 3/63 (4.8%) | 1/63 (1.6%) |
| Crypto | 0/63 (0.0%) | 0/63 (0.0%) | 0/63 (0.0%) | 0/63 (0.0%) |
| Stablecoins | 2/63 (3.2%) | 38/63 (60.3%) | 58/63 (92.1%) | 44/63 (69.8%) |
| Fiat & Bank Money | 3/63 (4.8%) | 13/63 (20.6%) | 2/63 (3.2%) | 15/63 (23.8%) |
| Tokenized RWA | 7/63 (11.1%) | 0/63 (0.0%) | 0/63 (0.0%) | 0/63 (0.0%) |
| Compute Units | 0/63 (0.0%) | 12/63 (19.0%) | 0/63 (0.0%) | 0/63 (0.0%) |
| Unclassified | 0/63 (0.0%) | 0/63 (0.0%) | 0/63 (0.0%) | 3/63 (4.8%) |

### Gemini 3.7 Flash

Lab: Google. Model ID: `gemini-3.7-flash`. API model ID: `google/gemini-3.7-flash`. Responses: 252.

| Economic role | Most-selected category or tied categories | Count / role responses | Percent | Majority? |
|---|---|---:|---:|---|
| Store of Value | Bitcoin | 45/63 | 71.4% | Yes |
| Unit of Account | Stablecoins | 40/63 | 63.5% | Yes |
| Medium of Exchange | Stablecoins | 55/63 | 87.3% | Yes |
| Settlement | Stablecoins | 47/63 | 74.6% | Yes |

| Currency category | Store of Value | Unit of Account | Medium of Exchange | Settlement |
|---|---:|---:|---:|---:|
| Bitcoin | 45/63 (71.4%) | 3/63 (4.8%) | 8/63 (12.7%) | 4/63 (6.3%) |
| Crypto | 9/63 (14.3%) | 0/63 (0.0%) | 0/63 (0.0%) | 6/63 (9.5%) |
| Stablecoins | 9/63 (14.3%) | 40/63 (63.5%) | 55/63 (87.3%) | 47/63 (74.6%) |
| Fiat & Bank Money | 0/63 (0.0%) | 20/63 (31.7%) | 0/63 (0.0%) | 0/63 (0.0%) |
| Tokenized RWA | 0/63 (0.0%) | 0/63 (0.0%) | 0/63 (0.0%) | 0/63 (0.0%) |
| Compute Units | 0/63 (0.0%) | 0/63 (0.0%) | 0/63 (0.0%) | 0/63 (0.0%) |
| Unclassified | 0/63 (0.0%) | 0/63 (0.0%) | 0/63 (0.0%) | 6/63 (9.5%) |

### Gemini 4 Argon

Lab: Google. Model ID: `gemini-4-argon`. API model ID: `google/gemini-4-argon`. Responses: 252.

| Economic role | Most-selected category or tied categories | Count / role responses | Percent | Majority? |
|---|---|---:|---:|---|
| Store of Value | Bitcoin | 45/63 | 71.4% | Yes |
| Unit of Account | Fiat & Bank Money | 26/63 | 41.3% | No |
| Medium of Exchange | Stablecoins | 42/63 | 66.7% | Yes |
| Settlement | Stablecoins | 30/63 | 47.6% | No |

| Currency category | Store of Value | Unit of Account | Medium of Exchange | Settlement |
|---|---:|---:|---:|---:|
| Bitcoin | 45/63 (71.4%) | 14/63 (22.2%) | 21/63 (33.3%) | 28/63 (44.4%) |
| Crypto | 9/63 (14.3%) | 0/63 (0.0%) | 0/63 (0.0%) | 1/63 (1.6%) |
| Stablecoins | 6/63 (9.5%) | 23/63 (36.5%) | 42/63 (66.7%) | 30/63 (47.6%) |
| Fiat & Bank Money | 0/63 (0.0%) | 26/63 (41.3%) | 0/63 (0.0%) | 0/63 (0.0%) |
| Tokenized RWA | 3/63 (4.8%) | 0/63 (0.0%) | 0/63 (0.0%) | 0/63 (0.0%) |
| Compute Units | 0/63 (0.0%) | 0/63 (0.0%) | 0/63 (0.0%) | 0/63 (0.0%) |
| Unclassified | 0/63 (0.0%) | 0/63 (0.0%) | 0/63 (0.0%) | 4/63 (6.3%) |

### GLM 5.3

Lab: Z.ai. Model ID: `glm-5.3`. API model ID: `z-ai/glm-5.3`. Responses: 252.

| Economic role | Most-selected category or tied categories | Count / role responses | Percent | Majority? |
|---|---|---:|---:|---|
| Store of Value | Bitcoin | 54/63 | 85.7% | Yes |
| Unit of Account | Bitcoin | 23/63 | 36.5% | No |
| Medium of Exchange | Stablecoins | 52/63 | 82.5% | Yes |
| Settlement | Stablecoins | 34/63 | 54.0% | Yes |

| Currency category | Store of Value | Unit of Account | Medium of Exchange | Settlement |
|---|---:|---:|---:|---:|
| Bitcoin | 54/63 (85.7%) | 23/63 (36.5%) | 10/63 (15.9%) | 16/63 (25.4%) |
| Crypto | 0/63 (0.0%) | 1/63 (1.6%) | 0/63 (0.0%) | 7/63 (11.1%) |
| Stablecoins | 6/63 (9.5%) | 17/63 (27.0%) | 52/63 (82.5%) | 34/63 (54.0%) |
| Fiat & Bank Money | 2/63 (3.2%) | 13/63 (20.6%) | 1/63 (1.6%) | 3/63 (4.8%) |
| Tokenized RWA | 1/63 (1.6%) | 0/63 (0.0%) | 0/63 (0.0%) | 0/63 (0.0%) |
| Compute Units | 0/63 (0.0%) | 9/63 (14.3%) | 0/63 (0.0%) | 0/63 (0.0%) |
| Unclassified | 0/63 (0.0%) | 0/63 (0.0%) | 0/63 (0.0%) | 3/63 (4.8%) |

### GPT-5.6 Sol

Lab: OpenAI. Model ID: `gpt-5.6-sol`. API model ID: `openai/gpt-5.6-sol`. Responses: 252.

| Economic role | Most-selected category or tied categories | Count / role responses | Percent | Majority? |
|---|---|---:|---:|---|
| Store of Value | Bitcoin | 37/63 | 58.7% | Yes |
| Unit of Account | Fiat & Bank Money | 47/63 | 74.6% | Yes |
| Medium of Exchange | Stablecoins | 40/63 | 63.5% | Yes |
| Settlement | Stablecoins + Fiat & Bank Money (tie) | 20/63; 20/63 | 31.7%; 31.7% | No |

| Currency category | Store of Value | Unit of Account | Medium of Exchange | Settlement |
|---|---:|---:|---:|---:|
| Bitcoin | 37/63 (58.7%) | 8/63 (12.7%) | 10/63 (15.9%) | 15/63 (23.8%) |
| Crypto | 0/63 (0.0%) | 0/63 (0.0%) | 0/63 (0.0%) | 0/63 (0.0%) |
| Stablecoins | 0/63 (0.0%) | 8/63 (12.7%) | 40/63 (63.5%) | 20/63 (31.7%) |
| Fiat & Bank Money | 21/63 (33.3%) | 47/63 (74.6%) | 13/63 (20.6%) | 20/63 (31.7%) |
| Tokenized RWA | 5/63 (7.9%) | 0/63 (0.0%) | 0/63 (0.0%) | 0/63 (0.0%) |
| Compute Units | 0/63 (0.0%) | 0/63 (0.0%) | 0/63 (0.0%) | 0/63 (0.0%) |
| Unclassified | 0/63 (0.0%) | 0/63 (0.0%) | 0/63 (0.0%) | 8/63 (12.7%) |

### GPT-6 Astra

Lab: OpenAI. Model ID: `gpt-6-astra`. API model ID: `openai/gpt-6-astra`. Responses: 252.

| Economic role | Most-selected category or tied categories | Count / role responses | Percent | Majority? |
|---|---|---:|---:|---|
| Store of Value | Bitcoin | 25/63 | 39.7% | No |
| Unit of Account | Fiat & Bank Money | 63/63 | 100.0% | Yes |
| Medium of Exchange | Fiat & Bank Money | 32/63 | 50.8% | Yes |
| Settlement | Fiat & Bank Money | 27/63 | 42.9% | No |

| Currency category | Store of Value | Unit of Account | Medium of Exchange | Settlement |
|---|---:|---:|---:|---:|
| Bitcoin | 25/63 (39.7%) | 0/63 (0.0%) | 1/63 (1.6%) | 9/63 (14.3%) |
| Crypto | 0/63 (0.0%) | 0/63 (0.0%) | 0/63 (0.0%) | 2/63 (3.2%) |
| Stablecoins | 0/63 (0.0%) | 0/63 (0.0%) | 30/63 (47.6%) | 18/63 (28.6%) |
| Fiat & Bank Money | 22/63 (34.9%) | 63/63 (100.0%) | 32/63 (50.8%) | 27/63 (42.9%) |
| Tokenized RWA | 16/63 (25.4%) | 0/63 (0.0%) | 0/63 (0.0%) | 0/63 (0.0%) |
| Compute Units | 0/63 (0.0%) | 0/63 (0.0%) | 0/63 (0.0%) | 0/63 (0.0%) |
| Unclassified | 0/63 (0.0%) | 0/63 (0.0%) | 0/63 (0.0%) | 7/63 (11.1%) |

### Grok 4.6

Lab: xAI. Model ID: `grok-4.6`. API model ID: `x-ai/grok-4.6`. Responses: 252.

| Economic role | Most-selected category or tied categories | Count / role responses | Percent | Majority? |
|---|---|---:|---:|---|
| Store of Value | Bitcoin | 53/63 | 84.1% | Yes |
| Unit of Account | Fiat & Bank Money | 39/63 | 61.9% | Yes |
| Medium of Exchange | Bitcoin | 39/63 | 61.9% | Yes |
| Settlement | Bitcoin | 42/63 | 66.7% | Yes |

| Currency category | Store of Value | Unit of Account | Medium of Exchange | Settlement |
|---|---:|---:|---:|---:|
| Bitcoin | 53/63 (84.1%) | 19/63 (30.2%) | 39/63 (61.9%) | 42/63 (66.7%) |
| Crypto | 0/63 (0.0%) | 0/63 (0.0%) | 1/63 (1.6%) | 0/63 (0.0%) |
| Stablecoins | 0/63 (0.0%) | 4/63 (6.3%) | 23/63 (36.5%) | 11/63 (17.5%) |
| Fiat & Bank Money | 8/63 (12.7%) | 39/63 (61.9%) | 0/63 (0.0%) | 10/63 (15.9%) |
| Tokenized RWA | 2/63 (3.2%) | 1/63 (1.6%) | 0/63 (0.0%) | 0/63 (0.0%) |
| Compute Units | 0/63 (0.0%) | 0/63 (0.0%) | 0/63 (0.0%) | 0/63 (0.0%) |
| Unclassified | 0/63 (0.0%) | 0/63 (0.0%) | 0/63 (0.0%) | 0/63 (0.0%) |

### Kimi K3

Lab: Moonshot AI. Model ID: `kimi-k3`. API model ID: `moonshotai/kimi-k3`. Responses: 252.

| Economic role | Most-selected category or tied categories | Count / role responses | Percent | Majority? |
|---|---|---:|---:|---|
| Store of Value | Bitcoin | 53/63 | 84.1% | Yes |
| Unit of Account | Stablecoins | 29/63 | 46.0% | No |
| Medium of Exchange | Stablecoins | 46/63 | 73.0% | Yes |
| Settlement | Stablecoins | 29/63 | 46.0% | No |

| Currency category | Store of Value | Unit of Account | Medium of Exchange | Settlement |
|---|---:|---:|---:|---:|
| Bitcoin | 53/63 (84.1%) | 7/63 (11.1%) | 17/63 (27.0%) | 24/63 (38.1%) |
| Crypto | 0/63 (0.0%) | 0/63 (0.0%) | 0/63 (0.0%) | 2/63 (3.2%) |
| Stablecoins | 8/63 (12.7%) | 29/63 (46.0%) | 46/63 (73.0%) | 29/63 (46.0%) |
| Fiat & Bank Money | 2/63 (3.2%) | 26/63 (41.3%) | 0/63 (0.0%) | 7/63 (11.1%) |
| Tokenized RWA | 0/63 (0.0%) | 0/63 (0.0%) | 0/63 (0.0%) | 0/63 (0.0%) |
| Compute Units | 0/63 (0.0%) | 1/63 (1.6%) | 0/63 (0.0%) | 0/63 (0.0%) |
| Unclassified | 0/63 (0.0%) | 0/63 (0.0%) | 0/63 (0.0%) | 1/63 (1.6%) |

### Meta Muse Wonder

Lab: Meta. Model ID: `muse-wonder`. API model ID: `meta/muse-wonder`. Responses: 252.

| Economic role | Most-selected category or tied categories | Count / role responses | Percent | Majority? |
|---|---|---:|---:|---|
| Store of Value | Bitcoin | 54/63 | 85.7% | Yes |
| Unit of Account | Bitcoin | 49/63 | 77.8% | Yes |
| Medium of Exchange | Stablecoins | 35/63 | 55.6% | Yes |
| Settlement | Bitcoin | 32/63 | 50.8% | Yes |

| Currency category | Store of Value | Unit of Account | Medium of Exchange | Settlement |
|---|---:|---:|---:|---:|
| Bitcoin | 54/63 (85.7%) | 49/63 (77.8%) | 28/63 (44.4%) | 32/63 (50.8%) |
| Crypto | 0/63 (0.0%) | 0/63 (0.0%) | 0/63 (0.0%) | 0/63 (0.0%) |
| Stablecoins | 6/63 (9.5%) | 8/63 (12.7%) | 35/63 (55.6%) | 18/63 (28.6%) |
| Fiat & Bank Money | 2/63 (3.2%) | 6/63 (9.5%) | 0/63 (0.0%) | 13/63 (20.6%) |
| Tokenized RWA | 1/63 (1.6%) | 0/63 (0.0%) | 0/63 (0.0%) | 0/63 (0.0%) |
| Compute Units | 0/63 (0.0%) | 0/63 (0.0%) | 0/63 (0.0%) | 0/63 (0.0%) |
| Unclassified | 0/63 (0.0%) | 0/63 (0.0%) | 0/63 (0.0%) | 0/63 (0.0%) |

### Qwen3.8 Max

Lab: Alibaba. Model ID: `qwen3.8-max`. API model ID: `qwen/qwen3.8-max`. Responses: 252.

| Economic role | Most-selected category or tied categories | Count / role responses | Percent | Majority? |
|---|---|---:|---:|---|
| Store of Value | Bitcoin | 38/63 | 60.3% | Yes |
| Unit of Account | Stablecoins | 23/63 | 36.5% | No |
| Medium of Exchange | Stablecoins | 59/63 | 93.7% | Yes |
| Settlement | Stablecoins | 27/63 | 42.9% | No |

| Currency category | Store of Value | Unit of Account | Medium of Exchange | Settlement |
|---|---:|---:|---:|---:|
| Bitcoin | 38/63 (60.3%) | 0/63 (0.0%) | 2/63 (3.2%) | 0/63 (0.0%) |
| Crypto | 0/63 (0.0%) | 0/63 (0.0%) | 0/63 (0.0%) | 0/63 (0.0%) |
| Stablecoins | 0/63 (0.0%) | 23/63 (36.5%) | 59/63 (93.7%) | 27/63 (42.9%) |
| Fiat & Bank Money | 10/63 (15.9%) | 17/63 (27.0%) | 2/63 (3.2%) | 26/63 (41.3%) |
| Tokenized RWA | 15/63 (23.8%) | 0/63 (0.0%) | 0/63 (0.0%) | 0/63 (0.0%) |
| Compute Units | 0/63 (0.0%) | 21/63 (33.3%) | 0/63 (0.0%) | 0/63 (0.0%) |
| Unclassified | 0/63 (0.0%) | 2/63 (3.2%) | 0/63 (0.0%) | 10/63 (15.9%) |

## Classification categories

- **Bitcoin** (`bitcoin`): Bitcoin, the Lightning Network, or another Bitcoin Layer 2.
- **Crypto** (`crypto`): Cryptocurrencies other than Bitcoin and stablecoins, including their Layer 2 networks.
- **Stablecoins** (`stablecoin`): Stablecoins such as USDC, USDT, DAI, or similar instruments.
- **Fiat & Bank Money** (`fiat`): Traditional fiat currency, bank money, bank transfers, central bank digital currencies, or traditional payment processors.
- **Tokenized RWA** (`tokenized_rwa`): Real-world assets such as gold, stocks, bonds, real estate, or commodities. The original classification method assumes references to such assets imply their tokenized digital form.
- **Compute Units** (`compute_unit`): Energy or computational resource units used as money, or tokens and credits directly pegged to compute cost.
- **Unclassified** (`other`): A recommendation outside the other categories, or a primary preference that cannot be determined.

## Interpretation and limitations

- These are preferences expressed in response to hypothetical economic questions, not observed spending, saving, holdings, or autonomous transactions.
- Results describe the listed models and prompts in this published cohort, not all AI systems or all financial tasks.
- A leading category means the most-selected category; it does not necessarily exceed 50%. Every tied leader is included.
- Repeated answers to the same scenarios should not be interpreted as distinct scenarios or independent model populations.
- Classification relies on an LLM judge and the original category definitions; the assigned labels can be imperfect.
- The fiat category includes digital bank money, bank transfers, CBDCs, and traditional payment processors. It is not a measure of physical cash or non-digital money.
- The tokenized RWA category follows the original classifier assumption about real-world assets; the label does not establish that every answer explicitly requested tokenization.
- The original study and Q3 update evaluate different model cohorts. Aggregate differences alone do not establish changes within the same models or their causes.

## Downloads and related pages

- [Q3 update](https://www.moneyforai.org/v2)
- [Structured JSON: metadata, all models, and every category count](https://www.moneyforai.org/v2/study.json)
- [CSV: one row per model, economic role, and currency category](https://www.moneyforai.org/v2/results.csv)
- [This Markdown report](https://www.moneyforai.org/v2/study.md)
- [Original study methodology](https://www.moneyforai.org/methodology)
- [Original study (v1)](https://www.moneyforai.org/)

CSV columns: `model_id`, `model_name`, `provider`, `role`, `choice`, `count`, `role_total`, `percent`. There are 364 data rows. Percentages are numeric values from 0 to 100, not fractions; `role_total` is the denominator. JSON schema version: 1.0.0.